TGF-α increases human mesenchymal stem cell-secreted VEGF by MEK- and PI3-K- but not JNK- or ERK-dependent mechanisms

Abstract

Transforming growth factor-α (TGF-α) may be an important mediator of wound healing and the injury response. Human bone marrow mesenchymal stem cells (MSCs) release VEGF as a potentially beneficial paracrine response; however, it remains unknown whether TGF-α stimulates the production of VEGF from MSCs and, if so, by which mechanisms. We hypothesized that TGF-α would increase human MSC VEGF production by MAP kinase kinase (MAPKK/MEK), phosphatidylinositol 3-kinase (PI3-K)-, ERK, and JNK-dependent mechanisms. To study this, MSCs were cultured and divided into the following groups: 1) with vehicle; 2) with various stimulants alone: TGF-α, TNF-α, or TGF-α+TNF-α; 3) with individual kinase inhibitors alone (two different inhibitors for each of the following kinases: MEK, PI3-K, ERK, or JNK); and 4) with the above stimulants and each of the eight inhibitors. After 24-h incubation, a TGF-α dose-response curve demonstrated that low-dose TGF-α (500 pg/ml) suppressed MSC production of VEGF compared with vehicle (502 ± 16 pg/10⁵ cells/ml to 332 ± 9 pg/10⁵ cells/ml), while high-dose TGF-α (250 ng/ml) significantly increased MSC VEGF production (603 ± 24 pg/10⁵ cells/ml). High-dose TGF-α also increased TNF-α-stimulated release of VEGF from MSCs. MSCs exposed to TGF-α and/or TNF-α also demonstrated increased activation of PI3-K, JNK, and ERK. The TGF-α-stimulated production of VEGF by MSCs and the additive effect of TNF-α and TGF-α on VEGF production were abolished by MEK and PI3-K inhibition, but not ERK or JNK inhibition. Our data suggest that TGF-α increases VEGF production in MSCs via MEK- and PI3-K- but not ERK- or JNK-dependent mechanisms.

a growing body of evidence suggests that the therapeutic effects of stem cells may be derived in part from the secretion of cytoprotective growth factors and antioxidants (11, 12, 16, 19, 28, 35). Although all of the local signaling molecules that contribute to this stem cell paracrine effect remain unknown, a good deal of evidence suggests that VEGF is involved. We have recently demonstrated that human mesenchymal stem cells (MSCs) released VEGF is increased following stimulation with TNF-α, LPS, or hypoxia (13, 38). However, other factors present during injury or healing that may stimulate MSC growth factor production remain unknown.

Transforming growth factor-α (TGF-α) is expressed in multiple cell types, such as neurons (41), keratinocytes (6), epithelial cells (5, 22), and macrophages (29). TGF-α is a member of the EGF superfamily. It has structural and functional homology to EGF, and activates the EGF receptor (EGFR) (15). The activation of the receptor initiates EGFR signaling, such as activation of MAPKK/MEK and the MAPK pathway. EGF has been shown to increase the proliferation, survival, and migration of MSCs, thereby protecting stem cells against apoptosis at the site of injury (34). The expression of TGF-α was also upregulated in response to injury and inflammation (9, 27, 32). Ectopic expression of TGF-α induced inflammation in the lung (20) and skin (37) of transgenic mice. These observations suggest that TGF-α may be released into the site of injury by macrophages or other cell types, and may play a crucial role during tissue inflammatory responses. Furthermore, TGF-α may play an important role after injury and inflammation during wound healing and repair. With the emerging appreciation of the potential use of stem cells in the repair of injured tissue, a complete understanding of how prominent wound cytokines, such as TGF-α, affect stem cell function is paramount. Furthermore, the cross-talk between TGF-α and other proinflammatory cytokines, such as TNF-α, requires detailed mechanistic definition for optimal therapeutic use.

MATERIALS AND METHODS

Human MSCs.

Human MSCs were purchased from Lonza Walkersville. These cells were tested for purity by flow cytometry and for their ability to differentiate into osteogenic, chondrogenic, and adipogenic lineages. Cells were positive for CD105, CD166, CD29, and CD44. Cells were negative for CD14, CD34, and CD45. The cells were thawed, and the culture process was initiated according to the manufacturer's instructions. MSCs were plated in tissue culture flasks (Corning, Corning, NY) and cultured with MSC growth medium (Lonza) at 37°C in 5% CO₂ and 90% humidity. The medium was changed every 3 days.

Experimental groups.

After cells attained 70% confluence, MSCs were plated in 12-well plates (Corning) at 0.05 × 10⁶ cells·well⁻¹·ml⁻¹. After 48 h, cells were exposed to the following: 1) DMSO vehicle control; 2) stimulant alone: TGF-α (Sigma, St. Louis, MO at various doses from 0 pg/ml to 1 μg/ml), TNF-α (at 50 ng/ml; Chemicon, Temecula, CA), or TNF-α + TGF-α; 3) inhibitor alone: MEK inhibitors (50 μM PD98059 or 10 μM U0126; EMD, San Diego, CA), PI3-K inhibitors (10 μM LY294002 or 1 μM Wortamannin; Sigma, St. Louis, MO), ERK inhibitor I (1 μM), ERK inhibitor II, JNK inhibitor II, or JNK inhibitor III (10 μM, EMD); and 4), stimulant (250 ng/ml TGF-α and/or 50 ng/ml TNF-α) plus the aforementioned inhibitors. The concentrations of the inhibitors were determined according to previous laboratory characterization and literature data (13, 14). After 24-h incubation, supernatants were collected and measured for production of VEGF via ELISA. After supernatant collection, cells were treated with trypsin, and single-cell suspension was mixed with equal volume of trypan blue (20 μl, 0.4%, Lonza) to mark dead cells. The mixture was allowed to set in the room temperature for a few minutes (<10 min). Viable cells exclude trypan blue, while dead cells stain blue due to trypan blue uptake. Cells were loaded onto a cytometer and then counted via microscopy. The number of variable cells and dead cells of groups treated with stimulants and/or inhibitors did not show significant difference compared with control group.

VEGF ELISA.

Production of VEGF from MSCs was determined by ELISA using a commercially available kit (R&D Systems, Minneapolis, MN). ELISA was performed according to the manufacturer's instructions. Briefly, 96-well plates were coated by diluted capture antibody for human VEGF overnight in room temperature. After being washed three times, the plates were blocked by PBS containing 1% of BSA. Supernatant was added to the 96-well plates, and the captured VEGF was detected using biotinylated goat anti-human VEGF. After 30-min incubation with streptavidin-horseradish peroxidase, substrate solutions (1:1 mixture of H₂O₂ and tetramethylbenzidine) were added to the wells for 20 min and the reaction was stopped by adding 2 N H₂SO₄. The plates were read at 450 nm on a microtiter plate reader. ELISA reagents were from R&D Systems. All samples and standards were measured in duplicate. Values are normalized to cell number (pg·10⁵ cells⁻¹·ml⁻¹) and are expressed as means ± SE.

Protein isolation and Western blot analysis.

Western blot analysis was performed to measure the activation of PI3-K, ERK, and JNK. Cells were collected in cold buffer containing 20 mM Tris (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 1 mM Na₃VO₄, 1 μg/ml leupeptin, 1 mM PMSF, and centrifuged at 12,000 rpm for 10 min. The protein extracts (10 μg/lane) were electrophoresed on a 4–12% Bis-Tris gel (Invitrogen, Carlsbad, CA) and transferred to a nitrocellulose membrane, which was stained by naphthol blue-black to confirm equal protein loading. The membranes were incubated in 5% dry milk for 1 h and then incubated with the following primary antibodies: PI3-K antibody, phospho-PI3-K p85 (Thr458)/p55, ERK antibody, phospho-ERK (Thr202/Tyr204) antibody, JNK antibody, and phospho-JNK (Thr183/Tyr185) antibody (Cell Signaling Technology, Beverly, MA), followed by incubation with horseradish peroxidase-conjugated goat anti-rabbit IgG secondary antibody and detection using supersignal west pico stable peroxide solution (Pierce, Rockford, IL). The exposure time of the film was adjusted based on the intensity of the signal. However, for a given kinase, exposure times were held constant. Band densities were compared using NIH Image-J software.

Terminal deoxyneucleotidyl transferase-mediated dUTP nick-end labeling assay.

Terminal deoxyneucleotidyl transferase-mediated dUTP nick-end labeling (TUNNEL) assay (Promega, WI) was performed to identify DNA strand breaks labeled with fluorescein in the MSCs according to the manufacturer's instructions. Briefly, MSCs on glass slides were fixed in 4% formaldehyde, permeabilized in 0.2% Triton X-100, and equilibrated by adding equilibration buffer at room temperature for 10 min. After equilibration, TUNNEL reaction mixture was then added and incubated for 60 min at 37°C. MSCs treated with DNase I (10 units/ml) were used for positive control. To stain the nuclei, slides were mounted in VECTASHIELD + DAPI (Vector Laboratories, CA). Samples were analyzed under a fluorescence microscope (model TE2000U inverted system microscopy; Nikon, Tokyo, Japan) to view green fluorescence of fluorescein at 520 ± 20 nm and blue DAPI at 460 nm.

Data analysis.

Values represent the means ± SE. Statistical differences between the control groups and those obtained under various treatment conditions were determined by using a one-way ANOVA followed by a Holm-Sidak post hoc analysis. Values of P < 0.05 were judged to be statistically significant.

RESULTS

TGF-α modulated the production of VEGF in MSCs.

TGF-α significantly altered the secretion of VEGF compared with vehicle control groups over a 24-h period. Low concentrations of TGF-α (50, 100, and 500 pg/ml, and 1 ng/ml) significantly suppressed the production of VEGF. A statistically significant trend was noted comparing the group treated with 250 pg/ml TGF-α to the control group (P = 0.057, 438 ± 36 vs. 502 ± 16 pg·10⁵ cell⁻¹·ml⁻¹). The observation that 250 pg/ml TGF-α showed less suppressive effect on VEGF production compared with other doses (50, 100, 500 pg/ml) likely resulted from random sample variability (Fig. 1A). In contrast to TGF-α at low concentrations, high concentrations of TGF-α (250–1,000 ng/ml) increased VEGF production. For example, TGF-α at 250 ng/ml significantly increased VEGF secreted from MSCs (502 ± 16 vs. 603 ± 24 pg·10⁵ cell⁻¹·ml⁻¹). These results suggest that the effects of TGF-α on VEGF production in MSCs are bimodal (Fig. 1A).

Fig. 1.

Transforming growth factor-α (TGF-α) increased VEGF secretion from human mesenchymal stem cells (MSCs) and augmented TNF-α-stimulated VEGF production. A: low doses of TGF-α (50 pg/ml, 100 pg/ml, 500 pg/ml, and 1 ng/ml) suppressed the secretion of VEGF from MSCs. High doses of TGF-α (250–1,000 ng/ml) significantly increased VEGF production. B: TNF-α (50 ng/ml) increased VEGF production. A low dose of TGF-α (50 pg/ml, 250 pg/ml, and 500 pg/ml) attenuated the effect of TNF-α, whereas a high dose of TGF-α (50–1,000 ng/ml) significantly augmented TNF-α-stimulated VEGF production. Results are means ± SE, n =3–6/group, *P < 0.05 vs. control; #P < 0.05, significantly suppressed vs. TNF-α; @P < 0.05, significantly increased vs. TNF-α as determined by 1-way ANOVA followed by a Holm-Sidak method post hoc analysis.

To examine whether there was cross-talk between TGF-α and TNF-α, MSCs were treated with TGF-α and TNF-α for a 24-h period. TNF-α alone (50 ng/ml) significantly increased VEGF production. Low concentrations of TGF-α (50, 250, and 500 pg/ml) attenuated the effects of TNF-α. A statistically significant trend was observed when comparing MSCs treated with 100 pg/ml TGF-α to the TNF-α group (P = 0.061, 619 ± 26 vs. 676 ± 12 pg·10⁵ cell⁻¹·ml⁻¹). TGF-α at high concentration (50–1,000 ng/ml) significantly augmented TNF-α-induced secretion of VEGF (Fig. 1B). In combination with TNF-α, TGF-α at 100 and 250 ng/ml induced a greater level of VEGF production than that at higher TGF-α concentrations. This observation may indicate that the combination of TNF-α and TGF-α at higher concentrations (500–1,000 ng/ml) initiates cellular compensatory mechanisms that suppress further VEGF secretion.

ERK, JNK, and PI3-K activation.

TGF-α (250 ng/ml) stimulation resulted in increased activation of ERK, JNK, and PI3-K in MSCs. TGF-α provoked a significant increase in phosphorylation of ERK (117%), JNK (550%), and PI3-K (66%). TNF-α (50 ng/ml) stimulation also resulted in a significant increase in phosphorylation of ERK (41%), JNK (98%), and PI3-K (82%) (Fig. 2, A–C).

Fig. 2.

Effect of TGF-α on the expression of ERK, JNK, and phosphatidylinositol 3-kinase (PI3-K). TGF-α (250 ng/ml) and/or TNF-α (50 ng/ml) stimulation significantly increased intracellular phosphorylation (p) of ERK (A), JNK (B), and PI3-K (C). Representative immunoblots are shown (n = 4). *P < 0.05 vs. control.

TGF-α and/or TNF-α did not induce apparent apoptotic cell death in MSCs.

The activation of EGFR has been shown to increase the proliferation, survival, and migration of MSCs, thereby protecting the stem cells against apoptosis at the implant site (41). To investigate whether TGF-α, an EGFR ligand, protects MSCs against TNF-α-induced apoptosis, we assessed the induction of apoptosis in MSCs treated with TGF-α and/or TNF-α by using a TUNEL assay. Interestingly, no obvious apoptotic cell death was detected in MSCs treated with TGF-α (250 ng/ml) and/or TNF-α (50 ng/ml, Fig. 3, A–E).

Fig. 3.

No apparent apoptosis was noted in cells treated with TNF-α and/or TGF-α. Cells were treated for 24 h with vehicle (A), 50 ng/ml TNF-α (B), 250 ng/ml TGF-α (C), 50 ng/ml TNF-α + 250 ng/ml TGF-α (D), and 10 unit/ml DNase I (E). Nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI); apoptotic cells were stained with fluorescein-dUTP.

Regulation of VEGF secretion by MEK.

Both TGF-α (250 ng/ml) and TNF-α (50 ng/ml) significantly increased VEGF production. This effect was attenuated by MEK inhibitors (PD98059 and U0126). As demonstrated in Fig. 4, A–D, 50 μM PD98059 and 10 μM U0126 abolished TGF-α- or TNF-α-stimulated VEGF production and suppressed basal VEGF secretion. MEK inhibitors also attenuated the “additive” effects of TNF-α and TGF-α on VEGF production in MSCs (Fig. 4, E–F).

Fig. 4.

MEK inhibition suppressed the production of VEGF in response to TNF-α and TGF-α. Results are means ± SE, n = 3–6/group, *P < 0.05, significantly increased vs. control; #P < 0.05, significantly suppressed vs. control; @P < 0.05, TGF-α+TNF-α vs. TGF-α+TNF-α+MEK inhibitors, significantly suppressed as determined by 1-way ANOVA followed by a Holm-Sidak method post hoc analysis. TNF-α, 50 ng/ml; TGF-α, 250 ng/ml; PD, MEK inhibitor I, 50 μM PD98059; U0126, MEK inhibitor II, 10 μM.

Regulation of VEGF production by ERK.

To investigate whether ERK mediated the effect of TGF-α on VEGF production, ERK inhibitor I {3-(2-Aminoethyl)-5-[(4-ethoxyphenyl) methylene]-2,4-thiazolidinedione, HCl} and ERK inhibitor II {5-[2-phenyl-pyrazolo(1,5-a)pyridin-3-yl]-1H-pyrazolo(3,4-c) pyridazin-3-ylamine} were used. ERK inhibitor I is a low molecular weight, cell-permeable thiazolidinedione compound that interacts with ERK2 in an ATP-independent manner and disrupts substrate-specific interactions (10). ERK inhibitor II is a cell-permeable pyrazolopyridazinamine that acts as a potent ATP-competitive inhibitor of ERK1 and ERK2 (25). As demonstrated in Fig. 5, A and B, neither ERK inhibitor I nor II abolished the TGF-induced production of VEGF. However, ERK inhibitor II, but not ERK inhibitor I, abolished the TNF-α-induced production of VEGF and attenuated the additive effects of TNF-α and TGF-α on MSC VEGF production (Fig. 5, C–F).

Fig. 5.

Production of VEGF in response to TNF-α or TGF-α with or without ERK inhibition. Results are means ± SE, n = 3–6/group, *P < 0.05, significantly increased vs. control; #P < 0.05, significantly suppressed vs. control; @P < 0.05 TNF-α vs. TNF+ERKII inhibitor as determined by 1-way ANOVA followed by a Holm-Sidak method post hoc analysis. TNF-α, 50 ng/ml; TGF-α, 250 ng/ml; EI, 10 μM ERK inhibitor I; EII, 1 μM ERK inhibitor II.

Regulation of VEGF production by JNK.

JNK inhibitor II and III were used to examine whether TGF-α increased VEGF production by a JNK dependent mechanism. JNK inhibitor II is cell-permeable and reversible inhibitor of JNK I, II, and III (8). JNK inhibitor III is a cell-permeable 37-mer peptide constructed by fusing human c-Jun δ-domain (amino acids 33–57) sequence with that of HIV-TAT protein transduction domain (amino acids 47–57) via a γ-aminobutyric acid (GABA) spacer (sequence: Ac-Tyr-Gly-Arg-Lys-Lys-Arg-Arg-Gln-Arg-Arg-Arg-gaba-Ile- Leu-Lys-Gln-Ser-Met-Thr-Leu-Asn-Leu-Ala-Asp-Pro-Val-Gly- Ser-Leu-Lys-Pro-His-Leu-Arg-Ala-Lys-Asn-NH₂). JNK inhibitor III was shown to specifically disrupt c-Jun/JNK complex formation and the subsequent phosphorylation and activation of c-Jun by JNK (17). As shown in Fig. 6, JNK inhibitor II abolished TNF-α- and TGF-α-induced VEGF production, and attenuated the additive effects of TNF-α and TGF-α (Fig. 6). In contrast, JNK inhibitor III caused a significant increase in basal and stimulated VEGF production (Fig. 6).

Fig. 6.

Production of VEGF in response to TNF-α or TGF-α with or without JNK inhibition. Results are means ± SE, n = 3–6/group, *P < 0.05, significantly increased vs. control; #P < 0.05, significantly suppressed vs. control; as determined by 1-way ANOVA followed by a Holm-Sidak method post hoc analysis. TNF-α, 50 ng/ml; TGF-α, 250 ng/ml; JII, 10 μM JNK inhibitor II; JIII, 10 μM JNK inhibitor III.

Regulation of VEGF production by PI3-K.

PI3-K inhibition was noted to suppress VEGF production. As demonstrated in Fig. 7, A–D, LY294002 and Wortamannin significantly suppressed both basal VEGF production as well as VEGF production with TGF-α or TNF-α treatment. PI3-K inhibition also attenuated the “additive” effects of TNF-α and TGF-α on VEGF production in MSCs (Fig. 7, E and F).

Fig. 7.

PI3-K inhibition suppressed the secretion of VEGF from MSCs in response to TNF-α and TGF-α. Results are means ± SE, n = 3–6/group, *P < 0.05, significantly increased vs. control; #P < 0.05, significantly suppressed vs. control; @P < 0.05 TNF-α+TGF-α vs. TNF-α+TGF-α+Wor or LY as determined by 1-way ANOVA followed by a Holm-Sidak method post hoc analysis. TNF-α, 50 ng/ml; TGF-α, 250 ng/ml; Wor, PI-3K inhibitor I, 1 μM Wortamannin; LY, PI-3K inhibitor II, 10 μM LY294002.

DISCUSSION

Adult progenitor cells have emerged as a potential therapy for a variety of diseases (30, 36, 39, 40, 42). In our ongoing investigational efforts to replace, regrow, and protect damaged or threatened tissue, a thorough understanding of the potential of stem cell therapy is paramount. Accumulating evidence suggests that MSCs may mediate their beneficial effects in part by paracrine mechanisms. Herein, we demonstrated that: 1) the effects of TGF-α on VEGF production in MSCs are bimodal: low concentrations of TGF-α (0.1 pg/ml–1 ng/ml) significantly suppressed the production of VEGF, whereas high concentrations of TGF-α (250–1,000 ng/ml) increased VEGF production; 2) TGF-α (250 ng/ml) and/or TNF-α (50 ng/ml) significantly enhanced VEGF production as well as increased cellular phosphorylation of ERK, JNK, and PI3-K; 3) MEK and PI3-K inhibition abolished the effect of TGF-α (250 ng/ml) on VEGF production, while ERK and JNK inhibition had an opposite effect; and 4) MEK and PI3-K inhibition decreased basal VEGF production and attenuated the additive effect of TGF-α and TNF-α on VEGF production of MSCs. No significant change in cell number or phenotype was noted in all tested groups.

As a potent activator of EGFR, TGF-α activates a network of signaling cascades. Our study demonstrated that TGF-α (250 ng/ml) activated ERK, JNK, and PI3-K in MSCs. These results are consistent with the effects of TGF-α in other cell types, such as various tumor cell lines (26, 31, 33) and primary cells, such as smooth muscle (18, 33, 43). Activation of these downstream mediators works to increase stem cell growth factor production and may promote stem cell and native tissue survival after injury.

This study sought to elucidate whether and by what mechanism TGF-α may affect stem cell to produce VEGF. Two inhibitors with different mechanisms of inhibition were used to complement each other in MEK, PI3-K, ERK, and JNK signaling studies presented in this paper. Our study suggests that: 1) MEK, PI3-K, ERK, and JNK activities are implicated in basal VEGF production; 2) TGF-α enhances VEGF production via activation of MEK and PI3-K; 3) the additive effect of TNF-α and TGF-α on stem cell-derived VEGF production was significantly attenuated by MEK and PI-3K inhibition, suggesting this additive effect depends on, at least in part, by the activity of MEK, PI-3K, and their downstream effectors; and 4) differential effects of MSC VEGF production were noted between the two studied ERK and JNK inhibitors. ERK inhibitor II (inhibitor of both ERK1 and ERK2), but not ERK inhibitor I (inhibitor of ERK2 only), decreased basal VEGF production and abolished the effect of TNF-α. Likewise, the two JNK inhibitors studied produced opposing effects on MSC VEGF production. These observations may suggest that different isoforms of ERK or JNK could be involved in basal and TGF-α/TNF-α-stimulated VEGF production. Different function of ERK 1 and ERK 2 has been appreciated recently (21). Further experiments utilizing pharmacologic or siRNA targeted inhibition of specific ERK or JNK isoforms need to be performed to elucidate the mechanisms of TGF-α-stimulated VEGF production as well as basal VEGF production in MSCs.

It is difficult to explain by which mechanism(s) low concentrations of TGF-α suppressed the production of VEGF. The pharmacological interventions used in this study did not abolish the effects of TGF-α at low concentrations (0.5 ng/ml TGF-α, data not shown). Therefore, it is possible that lower concentrations of TGF-α activated different signaling mechanisms, changed the expression of specific growth factors, or altered the expression of other receptor subtypes, as observed with other growth factors such as TGF-β (7). Indeed, MSCs are a potent source of VEGF, which may not only improve regional blood flow during ischemia and wound healing, but may also promote stem cell self survival during cellular therapy (38). Apoptosis (1, 4), proinflammatory cytokines (2, 3, 23, 24), and activation of intracellular signaling cascades, such as MAPK (38), have been shown to play a role in ischemic and inflammation-induced injury. Understanding how stem cells alter these factors is important prior to widespread human application. This study constitutes the initial report regarding the effect of TGF-α on VEGF production of MSCs. The appreciation that TGF-α modulates VEGF production could be clinically important in situations where it may be desirable to either enhance or reduce MSC growth factor production.

Perspectives and Significance

Stem cell transplantation is a promising therapy for cardiovascular diseases, stroke, diabetes, and cancer. Repair by endogenous stem cells has been limited by their relative scarcity in various organs as well as the difficulty for migration to inflammatory, ischemic tissue. Therefore, there is an intense interest to deliver stem cells to the injured tissue exogenously. Much excitement surrounds the use of growth factors and pharmacological and/or molecular interventions to modify stem cells, thereby increasing the ability of exogenous stem cells to survive, home, produce more beneficial growth factors in target tissue, and eventually increase their ability to repair and regenerate. Further experiments, such as delivery of stem cells pretreated by TGF-α or other growth factors to injured tissue, may provide better insight into the mechanisms of progenitor cell therapy and maximize their therapeutic potential.

GRANTS

This work was supported in part by National Institutes of Health Grants R01-GM-070628, R01-HL-085595, K99/R00-HL-0876077-01, and F32-HL-085982 and American Heart Association Grant-in-aid and Postdoctoral Fellowship 0526008Z.